Hormone replacement therapy, hormone levels, and lipoprotein cholesterol concentrations in elderly women

Gynecology Hormone replacement therapy, hormone levels, and lipoprotein cholesterol concentrations in elderly women Annlia Paganini-Hill, PhD," Ruth Dworsky, PhD, ~ and Ronald M. Krauss, MD b

Laguna HiUs, Los Angeles, and Berkeley, California OßJECTIVE: Our purpose was to assess the re ationships of lipid and lipoprotein cholesterol levels to hormone replacement therapy and hormone levels in elderly women. STLIDY DI=SlGN: A sample of 292 postmenopausal women 55 to 99 years old (mean 76 years) was drawn from Leisure World Laguna Hills, Californ a, an upper-middle-class, white independent-living population. We compared 84 women receiving unopposed estrogen replacement therapy and 38 women taking combination hormone replacement therapy with 170 women who had never used hormone replacement therapy. Nonparametric tests for differences in lipid and lipoprotein cholesterol levels among groups and multiple stepwise regression models were used. RI=$ULTS: Estrogen users (with and without progestin) had Iower total and Iow-density lipoprotein cholesterol and higher high-density lipoprotein and high-density lipoprotein subfraction types 2, 2a, and 2b cholesterol leveis. High-density lipoprotein type 3 subfractions were Iower in combination hormone replacement therapy users büt higher in unoppesed estrogen users relative to nonusers. The conjugated equine estrogen dose was negatively correlated with total (p = 0.0009) and Iow-density lipoprotein cholestero[ (p < 0.0001) levels and positively correlated to high-density lipoprotein cholesterol (p = 0.002) and its subfractions. The medroxyprogesterone acetate dose showed no consistent effect on choiesterol levels. CONCLU$1ON: The associations found here reaffirm the significant role of estrogen replacement therapy on lipid and lipoprotein cholesterol levels and provide no evidence of a reduction in the beneficial effect of estrogen with the addition of a progestational agent to the replacement regimen. (AMJ OBSTETGYNECOL 1996;174:897-902.)

Key words: Estrogen, progestin, lipoproteins, cholesterol, triglycerides, h o r m o n e replacement therapy

Estrogens are an important category of medication for menopausal women. Surveys suggest that between 10% and 35% of postmenopausal women, depending on geographic location, are currently taking estrogens.l' 2 Use of specific agents has varied over the years. Unfil the middle From the Department of Preventive Medicine, Univet~ity of Southern Cali]brnia School of Medicine,~ and the Lawrence Berkeley National Laboratory, Department of Molecular and Nuclear Mea'icine, University of California, Berheley.~ Supported by grant No. CA32197 from the National CancerInstitute, grants No. HL18574 and No. HL33577 from the National Heart, Lung, and Blood Institute of the National Institutes of Health, and grants from the Wright Foundation and Wyeth-Aye~st Laboratories. Conducted in part at the Lawrence Berkelo, National Laboratory through the U«~. Department of Energy untier conlract No. DE-ACO376F00098. Receivedfor publication April 10, 1995; revisedJuly 12, 1995; accepled July 25, 1995. Reprint requests: Annlia Paganini-l+Sill, PhD, Unive~osityof Soutt~ern California School ofMedicine, 1721 Griffin Ave #200, Los Angeles, CA 90031. Copyright © 1996 óy Mosby-Year Book, Inc. 0002-9378/96 (~5.00+ 0 6/1/68221

1960s most of the drugs were synthetic estrogens that were graduaily replaced with natural estrogens, primarily conjugated equine estrogens (Premarin). The 1980s saw the introducüon of the regimen of estrogen cycled or continuously used with a progestational agent to reduce the risk of endometrial hyperplasia and carcinoma. Recent studies have shown that use ofestrogen replacement therapy by poslmenopausal women protects against coronary heart disease and stroke. 3 This effect may be due in part to alterations in serum lipoprotein levels among estrogen users. These changes include increased high-density lipoprotein (HDL) cholesterol and decreased Iow-density lipoprotein (LDL) cholesmrol levels. 4 Questions still remain, howevel, regarding the effect of combination estrogen-progestin therapy on lipids and lipoproteins. Clinical studies, offen of short duration, have shown that many progestins oppose the beneficial effects of estrogen alone, particularly changes in HDL cholesterol. If so, combination hormonal replacement therapy may increase the risk of coronary heart disease 897

898 Paganini-Hill, Dworsky, and

Krauss

March 1996

Ami Obstet Gynecol

T a b l e I. Characteristics o f study subjects by use o f h o r m o n e r e p l a c e m e n t t h e r a p y

No. Age (yr) (mean + SEM) Height (in) (mean _+SEM) Weight (lb) (mean ± SEM) Body mass (1000 • weight/height 2) (mean ± SEM) Exercise (hr/wk) (mean _+SEM) Alcohol (drinks/wk) (mean + SEM) Current smoker (%) Regular use (%) Analgesics Vitamin supplements Antihypertensive medication Thyroid medication

Never used HR T

Unopposed ER T

Combination H R T

170 77 _+0.5 63 + 0.2 134 _+1.6 33 -+ 0.4

84 76 + 0.6 63 _+0.3 132 + 2.5 33 + 0.5

38 71 + 1.0 63 + 0.5 137 _+3.4 34_+0.8

3.4 -+0.2 3.1 + 0.4 5.9

3.0 ± 0.4 2.7 ± 0.5 10.7

3.2+0.5 2.3 + 0.6 10.5

28.2 34.7 22.4 13.5

41.7 36.9 34.5 25.0

47.4 47.4 18.4 26.3

Significance p < 0.0001

p = 0.04 p = 0.06

HRT, Hormone replacement therapy; ERT, estrogen replacement therapy.

T a b l e II. M e a n _+SEM o f lipid a n d l i p o p r o t e i n c h o l e s t e r o l a n d h o r m o n e levels by use o f h o r m o n e replacement therapy

Triglycerides Total cholesterol LDL cholesterol HDL cholesterol HDLz HDL2a HDL2b HDL3 HDL3a HDL3b HDL~« Hormone levels* Progesterone Estrone Estradiol Steroid hormone-binding globulin

Never used H R T

Unopposed ERT

Com&'nation HRT

127 ± 4.5 232 ± 3.2 147 _+2.9 60 + 1.2 33 + 1.0 15 ± 0.3 18 _+0.8 28 + 0.4 17 + 0.3 7.7 _+0.1 3.3 ± 0.1

130 + 7.4 218 ± 4.0 124 -+ 3.7 68 -+ 1.8 39 + 1.7 18 -+ 0.6 21 -+ 1.2 30 + 0.7 18 ± 0.4 8.2 -+ 0.3 3.4 _+0.1

129 + 12.1 219 ± 6.4 128 -+ 5.9 69 + 3.1 41 ± 2.8 18 ± 0.8 24 ± 2.2 27 -+ 1.1 17 -+0.6 7.2 + 0.4 2.7 +- 0 9

30 ± 2.3 3.1 ± .14 0.64 + .03 1.5 _+.12

28 -+ 2.5 8.9 -+ 1.3 2.6 ± .85 2.3 ± .22

26 + 1.3 7.5 -+ .92 1.7 _+.28 2.8 + .22

Signißcance p = 0.02 p < 0.0001 p = 0.0001 p = 0.0003 p< 0.0001 p = 0.004 p = 0.01 p = 0.01 p = 0.03 p < 0.0001 p < 0.0001 p< 0.0001

HRT, Hormone replacement therapy; ERT, estrogen replacement therapy. *Hormone levels were measured on 101 individuals: 35 never users of hormone replacement therapy, 33 using unopposed estrogen replacement therapy, and 33 using combinadon hormone replacement therapy.

a n d o t h e r a t h e r o s c l e r o t i c c o m p l i c a t i o n s relative to use o f e s t r o g e n alone. To assess t h e r e l a t i o n s h i p o f h o r m o n e r e p l a c e m e n t t h e r a p y a n d h o r m o n e levels to lipid a n d l i p o p r o t e i n c h o lesterol levels, we c o n d u c t e d a cross-sectional study in a c o m m u n i t y o f o l d e r adults, w h o are at t h e g r e a t e s t risk o f c o r o n a r y h e a r t disease. We s t u d i e d a s a m p l e o f 292 w o m e n with a n average age of 76 years f r o m this i n d e p e n dent-living p o p u l a t i o n .

Material and methods Study subjects were selected f r o m t h e 13,979 particip a n t s in t h e Leisure W o r l d C o h o r t Study. We b e g a n this c o h o r t study in J u n e 1981 w h e n we m a i l e d a d e t a i l e d h e a l t h survey to all r e s i d e n t s o f Leisure W o r l d L a g u n a Hills, a n u p p e r m i d d l e class r e t i r e m e n t c o m m u n i t y n e a r

Los Angeles. T h e h e a l t h q u e s t i o n n a i r e r e q u e s t e d , a m o n g o t h e r data, i n f o r m a t i o n o n c e r t a i n p r i o r m e d i c a l diagnoses, i n c l u d i n g a n g i n a , acute myocardial infarction, hyp e r t e n s i o n , a n d diabetes, a n d m e n s t r u a l a n d r e p r o d u c tive history, i n c l u d i n g e s t r o g e n r e p l a c e m e n t therapy. We followed u p t h e c o h o r t for all hospital admissions to t h r e e hospitals serving t h e a r e a a n d for d e a t h s identified by local a n d state h e a l t h d e p a r t m e n t s a n d p e r f o r m e d a search o f t h e N a t i o n a l D e a t h I n d e x . In a d d i t i o n , we c o n d u c t e d t h r e e follow-up mailings in 1983, 1985, a n d 1992. F r o m t h e 8877 w o m e n w h o c o m p l e t e d t h e questionnaire, t h o s e w h o h a d a n a t u r a l m e n o p a u s e ; h a d n o history o f myocardial i n f a r c t i o n , angina, stroke, or d i a b e t e s o n t h e basis o f t h e original o r follow-up q u e s t i o n n a i r e s or hospital d i s c h a r g e diagnoses; h a d n e v e r u s e d h o r m o n e

Paganini-Hill, Dworsky, and Krauss 899

Volume 174, Number 3 AmJ Obstet Gynecol

Table III, Mean + SEM of lipid and lipoprotein cholesterol levels by dose of estrogen and progestogen

dose*

No. Triglycerides cholesterol

Unopposedestrogen 0.3mg 0.625 nag >0.9 nag

25 45 9

125±11.7 138 ± 11.4 125 ± 19.6

224±8.4 217 ± 5.2 218 ± 9.9

Estrogenplus medroxyprogesteroneacetate<-5mg

I2DL ] HDL

HDLz

134±7.3 65±3.335 ±3.2 I21±5.0 69±2.5 39 ±2.3 [19±10.6 76±6.2 45 ±4.8

HDL2~` HDL2o HDL3 ] HDLg~ [ HDLgz~ HDLg~ 16±1.0 19±2.3 2 9 ± 1 . 6 1 8 ± 0 . 9 8 . 3 ± 0 . 6 3 . 4 ± 0 , 3 18±0.8 21±1.7 3 0 ± 1 . 0 1 8 ± 0 . 6 8 . 4 ± 0 . 4 3 . 4 ± 0 . 2 19±1.5 26±3.7 3 1 ± 2 . 3 1 9 ± 1 . 7 8 . 2 ± 0 . 6 3 . 4 ± 0 . 4

123±11.0 71±5.5 47±11.1 18±3.4 29±8.3 24±2.3 15±1.96.2±0.8 2.5±0.5 120±7.4 76±4.1 48±5.0 20± [.1 28±4.1 29±1.4 18±0.8 7.6±0.6 2.7±&3 25 16 8.0 1.9 11 11 133 47 22

0.3mg 0.625 mg >-0.9 nag

5 9 l

105±32.2 110±11.9 472

215±8.6 2t8±8.6 274

0.3 nag 0.625 nag >-0.9 mg

6 9 1

147±20.6 128 ± 12.5 49

213_+14.5 142±16.5 58±8.7 31 ±7.5 227±13.8 126±14.8 75±5.0 45 ±4.5 48 132 61 61

36 64 1i

126 ± 10.0 133 ± 8.4 149±36.6

221 ± 6.4 218_+4.4 216 ± 12.6 p < 0.0001

Estrogenplus medroxyprogesteroneacetate 10 rag

16±1.7 15±6.0 27±1.9 17±1.1 7.3±0.6 2.4__+0.3 19±1.3 26±3.7 30±2.6 19±1.7 7.9±0.8 2.9±0.3 13 13 0 0 17 31

TOTALS

0.3 mg 0.625 nag >-0.9 mg

134±5.9 64±3.136±2.9 |21±4.2 70±2.141±1.9 115±10.2 72±5.7 43±4.4 ,p=0.0002

16±0.820±2.2 2 8 ± 1 . 2 1 7 ± 0 . 7 7 . 8 ± 0 . 5 3 . 0 ± 0 . 3 18±0.623±1.4 3 0 ± 0 . 8 1 8 ± 0 . 5 8 . 2 ± 0 . 3 3 . 2 ± 0 . 3 18±t.425±3.4 2.9±2.518±1.67.4±0.93.0±0.5

*Conjugated equine estrogen: Premarin and generic.

replacement therapy or were currently taking h o r m o n e replacement therapy at the follow-up in 1985; and still resided in the community in 1987 were identified (n = 1078). From these we obtained a sample of 292 white females, 57 to 99 years old, who gave informed written consent before the study. The Institutional Review Board of the University of Southern California School of Medicine approved the study. Height and weight were measured with light clothing and without shoes. Information was collected on current smoking status and alcohol consumption, hospitalizations in the past year, chronic disease history, and intake of prescription and over-the-counter medications consumed regularly (at least once aweek). Blood samples were obtained between 7 and 10 AMafter an overnight fast by yen±puncture in tubes containing 2.7 m m o l / L potassium ethylenediaminetetraacetic acid. Plasma was separated by centrifugal±on (1000g, 4 ° C), packed in wet ice, and sent by overnight shipment to the laboratory of R.M.K. Measurements of total cholesterol, HDL cholesterol, and triglyceride levels were performed by standardized Lipid Research Clinics techniques) LDL cholesterol levels were calculated by the following standard formula: LDL cholesterol = Total cholesterol - HDL cholesterol - (Triglycerides/5). HDL subfraction distributions were determined by nondenaturing gradient gel electrophoresis as described by Nichols et al. ~ Densitometrywas used to measure areas of five protein-stained HDL bands (HDL2~, HDL2b, HDLa~, HDL3b , and HDL3~ ). Estimates of the plasma concentrations of these HDL subfraclions were obtained by multiplying the percent of total area for each band by the total plasma HDL cholesterol concentration. Assays for progesterone, estrone, estradiol, and steroid h o r m o n e binding globulin were performed on samples

from 101 of these women: 35 nonusers of h o r m o n e replacement therapy, 33 users of unopposed estrogen replacement therapy, and 33 users of combination hormone replacement therapy. Methods for these assays have been previously described7 Differences in lipid and lipoprotein cholesterol and h o r m o n e levels among groups was tested with standard nonparametric measures: the Mann-Whitney U statistic for two groups and the Kruskal-Wallis Hstatistic for three or more groups. Multiple stepwise regression models were constructed with triglycerides, total cholesterol, LDL cholesterol, HDL cholesterol, and the HDL subfractions 2, 2a, 2b, 3, 3a, 3b, and 3c separately as the dependent variables with age (years); height (inches); weight (pounds); body mass (weight/height "2× 1000); smoking (current cigarettes per day); alcohol (drinks per week); exercise (hours per week); use of estrogen, progestin, hypertensive medication, thyroid medication, analgesics, multivitamin supplements, vitamin C and B supplements; and circulating levels of estrone, estradiol, progesterone, and steroid hormone binding globulin as explanatory variables. Variables were added one by one to the models until no remaining variable produced a significant F statistic (p < 0.01) that reflects the variable's contribution to the model if it is included. The reported p values from the multiple linear regression models indicate the statistical significance of a given independent variable after the effects of the other independent variables on the lipid or lipoprotein cholesterol levels are considered.

Results Table I describes the characteristics of the study subjects by use of h o r m o n e replacement therapy: women who had never used h o r m o n e replacement therapy

900

Paganini-Hill, Dworsky, and Krauss

March t996

AmJ ObstetGynecol

(n = 170), women who were taking unopposed estrogen replacement therapy (n = 84), and women who were taking combination hormone replacement therapy (n = 38). Women who were taking combination hormone replacement therapy were on average younger than the women in the other two groups. Height, weight, exercise, and alcohol consumption did not differ among the three groups. H o r m o n e replacement therapy users were more likely to be taking thyroid medication and analgesics and were somewhat more likely to be smokers than were women who had never used hormone replacement therapy. Table II gives the mean _+SEM concentrations of serum triglycerides, total cholesterol, and lipoprotein cholesterol fractions by use of hormone replacement therapy. Estrogen users (with and without progestin) had lower total cholesterol levels, lower LDL cholesterol levels, and higher HDL cholesterol levels; triglyceride levels were not statistically different. All these lipid levels were similar in estrogen replacement therapy users compared with those in combination hormone replacement therapy users. HDL~ subfractions (both HDL~~~ and HDI,gb) were significantly higher in hormone replacement therapy users than in nonusers. HDL~ subfractions tended to be lower in combination hormone replacement therapy users but higher in unopposed estrogen replacement therapy users relative to nonusers. Table II also gives the plasma hormone levels in the subsample of 101 women. Estrone, estradiol, and steroid hormone binding globulin were all significantly higher in h o r m o n e replacement therapy users relative to nonnsers. Of the 122 women taking estrogens in this study, 78 were taking Premarin, 33 generic brands of conjugated estrogens, 3 estropipate (Ogen), 3 ethinyl estradiol (Estinyl), 3 estradiol patch (Estraderm), 1 esterified estrogens and methyltestosterone (Estratest) and 1 estradiol (Estrace). All were taking oral preparations, except for 3 women using Premarin cremn and the 3 women using the patch. Of the 38 women taking progestational agents, 36 were using medrox)3orogesterone acetate and 2 norethindrone acetate. The prescription schedule varied among the women. The most frequent (n = 54) pattern was estrogen for 20 to 25 days a month; 25 of these women also took a progestational agent from 5 to 25 days a month. Other frequent patterns were estrogens 5 days per week (n = 22) with progestins on the same days (n = 6) or on the other days (n = 1), estrogens every day or every other day of the month ( n = 19) with progestins every day ( n = 3) or 10 days of the month ( n = 2), and estrogens 3 times a week ( n = 12). One woman had another cyclic pattern: estrogens on days 1 to 13 and progestins on days 13 to 23. The other 14 women took unopposed estrogens ranging Dom one to four times a week. Table III give the mean lipid and cholesterol levels for the women receiving hormone replacement therapy by

the dose of medication. There was no consistent effect of medroxyprogesterone acetate dose, but the conjugated equine estrogen dose was negatively correlated to total cholesterol (r=-0.20, p = 0.0009) and LDL cholesterol (r= -0.29, p < 0.0001) and positively related to HDL cholesterol (r= 0.19, p = 0.002) and its subfractions HDL 2 (r=0.21, p=0.0006), HDL2~~ (r=0.22, p=0.0002), HDLgb (r=0.18, p=0.0Ô2), and HDLsa (r=0.12, p = 0.05). In multiple regression modeling the dose of estrogen replacement therapy explained 4% of the variability of total cholesterol and 8% of the variability of LDL cholesterol; no other variable made a significant contribution after estrogen was controlled for. The HDL cholesterol level was independently related to estrogen dose (p = 0.005), body mass (p = 0.0001), and age (p = 0.006), which together explained 13 % of the variability. Estrogen dose, body mass (or weight), alcohol consumption, age, and n u m b e r of cigarettes were also related in multivariate analyses to the HDL subfractions. Triglyceride levels were associated with weight (p = 0.0001), which accounted for 5% of its variability.

Comment Although the women were not randomly assigned to the hormone replacement therapy treatment groups in this study, factors affecting lipid and lipoprotein levels such as weight, exercise, and alcohol consumption were evenly distributed among the groups. Age and smoking status have also been shown to affect lipoprotein cholesterot levels and were controlled for in this study. Although these are cross-sectional data and direct causal relations cannot be inferred from them, the results show the relationship belween hormone replacement therapy use and lipid and lipoprotein levels in a nonclinical setting. These data generally agree with previously published studies that postmenopausal women treated with estrogen replacement therapy have a decrease in total cholesterol, principally because of a decrease in LDL cholesterol, and an increase in HDL cholesterol and triglycerides.4 Bush and Miller4 have summarized the results of randomized and cross-over studies of estrogen use and lipid and lipoprotein levels in menopausal women. The fange of lipid and lipoprotein response reported in these studies is wide due to a variety of methodologic factors, including size of the study sample, duration of use within the study; and dose of estrogen. Adjusting 10 studies for sample size (ranging from 6 to 265) and duration ofstudy (ranging from 1 month to 1 year), they found that the adjusted percent change in lipid and lipoprotein after administration of equine estrogens was -9% for total cholesterol, +10% for HDL cholesterol, -6% for LDL cholesterol, and +20% for triglycerides. Several cross-sectional studies have examined the erz

Yblume 174, Number 3 AmJ Obstet Gyneco]

fect of estrogen replacement therapy, especially conjugated equine estrogens, on serum lipids and lipoprotein cholesterol levels. The Lipid Research Clinics of North America Study compared 370 nonmenstruating women 45 to 64 years old not taking estrogens with 239 others taking equine estrogens. ~ They found the median serum triglyceride level to be 26% higher, total cholesterol 4% higher, HDL cholesterol 10% higher, and LDL cholesterol 11% lower in estrogen users. All these differences were statistically significant (p < 0.0ä). Another large cross-sectional study, the Cardiovascuiar Heart Study, compared lipoprotein lipid concentrations in 2732 women >65 years old (median 72.5 ~mars).9 Estrogen users had 4% higher triglyceride levels, 23% higher HDL cholesterol levels, and 15% lower LDL cholesterol levels than nonusers did. These resutts are consistent with our findings in a similarly aged group o f w o m e n (median 76 years) of differences of +5% in triglyceride, +13% in HDL cholesterol, and -18% in LDL cholesLerol. Our study showing more favorable lipoprotein cholesterol profiles at all dose levels of conjngated equine estrogen agrees with the summary of Bush and Mitler4: with low dose (0.625 mg) the percent differences relative to nonusers in triglycerides, total cholesterol, LDL cholesterol, and HDL cholesterol levels were 5%, -6%, -18%, and 17%, respectively, in our study and 1 1 % , - 1 % , - 4 % , and 10%, respectively, in their summary report. At higher dose (>0.9 mg), the differences were 17%: -7%, -22%, and 20% in our smdy and 17%, -3%, -8%, and 14% for a dose of 1.25 mg in the summary report. In a double-blind placebo-controlled crossover trial, Walsh et al) ° treated 31 postmenopausal women with 0.625 mg conjugated equine estrogen for 3 months and w~th 1.25 mg for another 3 months. They found that, at a dose of 0.625 mg triglycerides, total cholesterol, LDL chotesterol, and HDL cholesterol changed 24%, -4%, -15%, and 16%, whereas at a dose of 1.25 mg these changes were 38~, -6%, -19%, and 18%. Others have found similar dose effects. 1' Thüs the evidence that unopposed oval estrogen therapy modifies lipid and lipoprotein levels in a dose-dependent manner that should protect against the development of coronary heart disease (e.g., HDL cholesterol levels increased and LDL cholesterol levels decreased) is strong and consistent. We had only six nonoral (creams or percutaneous abdominal patches) estrogen users. With this ,';mail number we were unable to detect any differences between oral and nonoral estrogen users. Several short-term randomized or cross-over studies have shown that adding a progesterone-derived progestin to estrogen replacement therapy decreased the beneilt of estrogen replacement therapy on HDL cholesterol levels. ~'-~~'In two studies' u' *:~with 10 mg of oral medroxyprogesterone acetate prescribed for 10 days in each 28day cycle, HDL cholesterol returned to baseline. One

Paganini-Hill, Dworsky, and Krauss 901

study found that 10 mg medroxyprogesterone acetate for 13 days each cycle reduced HDL cholesterol 17% be!ow that with unopposed estrogen and 7% below baseline values? 5Another study of medroxyprogesterone acetate '~ and a study of oral medrogesteron ~4 used 5 mg for 10 to 11 days and found smaller reductions in the benefit compared with estrogen replacement therapy atone. The recently completed 3-year Postmenopausal Estrogen/Progestin Interventions Trial of 875 postmenopausal women 45 to 64 years old found that women randomized to receive conjugated equine estrogen with or without medroxyprogesterone acetate or micronized progesterone had significantly greater increases in mean HDL cholesterol than the placebo group did. "~The average increases were simitar in women assigned to estrogen alone (7%) or estrogen with micronized progesterone (9%), and these were greater than in women assigned estrogen with cyclic or continuous medroxyprogesterone acetate (2%). Mean LDL chotesterol levels decreased (10% to 13%~ and triglyceride levels increased (6%) in women receiving all acuve treatments and differed slgnificanfly from placebo /-3% and +2%. respectively). In Leisure World we observed no differences in lipoprotein cholesterol ievels between unopposed estrogen replacement therapy or combination hormone replacement therapy, hut patterns of use of both estrogens and progestins varied greatly among the women. It is also possible that progestin-induced alterations of lipoprotems (notably HDLs) attenuate over time. Few studies have examined hormonal determinants of HDL subfractions. Several studies ~° ~~-~ä ~7-,_,oreported an estrogen-induced increase in HDL 2. In two studies the HDL~ subfraction was substantialty increased by oral estrogen therapy. ~°~ 2~ In another studv ~'2 treatment »~ith conjugated eswogens caused HDL 2 levels to rise quickty; HDLs levels gradually increased bm did not reach statistical significance. Our study supports an esrrogen-induced increase in HDL 2. Unopposed estrogen users also had higher HDL 3 Ievels, but the increase was not observed in combination h o r m o n e reptacement therapy users. In a short-term (3 months~ trial Ottosson et a!. '~ found that the cyclic addition of medroxyprogesterone acetate decreased HDL~ levels. Miller et al. 'ä found that cyclic medroxyprogesterone acetate decreased both HDL 2 and HDLs levels. Analyses of individual components within the major HDL subclasses indicated that both HDLu~~ and HDL._,b contributed to hormone-related elevations in H D L » whereas the higher HDL a in es~rogen users was primarily related to HDL3~. These results are generaily compatible with another recent analysis in a smaller group of women.~~ although in this latter report the estrogen-related increase in HDLub was of smaller magnitude than that of HDL2~. Although the potential significance of these HDL components to atherosclerosis risk has not

902

Paganini-Hill, Dworsky, and Krauss

b e e n established, some e v i d e n c e suggests t h a t c o r o n a r y h e a r t disease risk may b e h i g h e r w h e n HDLzb is dec r e a s e d relative to HDL3b a n d H D L s « z4 O u r results o f t h e effect of p o s t m e n o p a u s a l estrogenp r o g e s t i n use o n lipid a n d l i p o p r o t e i n profiles are similar to those o f a cross-sectional study o f 1057 w o m e n 50 to 79 years old in R a n c h o B e r n a r d o , Calif., by B a r r e t t - C o n n o r et al. z5 As in o u r study, n o d i f f e r e n c e s in lipid or l i p o p r o tein c h o l e s t e r o l levels were n o t e d in w o m e n receiving e s t r o g e n r e p l a c e m e n t t h e r a p y versus c o m b i n a d o n horm o n e r e p l a c e m e n t therapy. In b o t h studies the h o r m o n e r e p l a c e m e n t t h e r a p y users h a d u s e d h o r m o n e replacem e n t t h e r a p y for several years. T h e s e studies i n d i c a t e t h e i m p o r t a n c e o f m o n i t o r i n g lipid c h a n g e s in l o n g - t e r m users o f h o r m o n e r e p l a c e m e n t therapy. T h e lowering o f H D L c h o l e s t e r o l by p r o g e s t i n m a y b e time d e p e n d e n t with a less negative effect after p r o l o n g e d use. We t h a n k M a r y Beck a n d Beverly Ducey for t h e i r assist a n c e with data collection a n d L a u r a Holl a n d C h a r l o t t e B r o w n for lipid a n d l i p o p r o t e i n assays. We are i n d e b t e d to the r e s i d e n t s o f Leisure World, L a g u n a Hills, w h o s e c o o p e r a t i o n m a d e this work possible.

REFERENCES

1. Derby CA, Hume AL, Barbour MM, McPhillips JB, Lasater TM, Carleton RA. Correlates of postmenopausal estrogen use and trends through the 1980s in two southeastern New England communities. Am J Epidemiol 1993; 137:112535. 2. Barrett-Connor E, ga'itz-Silverstein D. Estrogen replacement therapy and cognitive function in older women.JAMA 1993; 269:2637-41. 3. Stampfer MJ, Colditz GA. Estrogen replacement therapy and coronary heart disease: a quandtative assessment of the epidemiologic evidence. Prev Med 1991;20:47-63. 4. Bush TL, Miller VT. Effects of pharmacologic agents used during menopause: impact on lipids and lipoproteins. In: Mishell DR, ed. Menopause: physiotogy and pharmacology. Chicago: Year Book Medical, 1987:187-208. 5. Lipid Research Clinics Program. Manual of laboratory operations: lipid and lipoprotein analysis. Washington, DC: U.S. Department of Health, Education, and We!fare, 1974. (Publication no NIH 75-628.) 6. Nichols AV, Krauss RM, Musliner TA. Nondenaturing polyacrylamide gradient gel electrophoresis. Methods Enzymol 1986;128:417-31. 7. Wu 'C-H, Lundy LE. Radioimmunoassay of plasma estrogens. Steroids 1971;18:91-111. 8. Wahl E Walden C, gampp R, et al. Effect ofestrogen/progestin potency on lipid/lipoprotein cholesterol. N Engl J Med 1983;308:862-7. 9. Ettinger WH, Wahl PW, Kuller LH, et al. Lipoprotein lipids in older people: results from the Cardiovascular Health Study. Circulation 1992;86:858-69. 10. Walsh BW, Schiff I, Rosner B, Greenberg L, Ravnikar V, Sacks FM. Effects of postmenopausal estrogen replacement

March 1996 Am,l Obstet Gynecol

11.

12.

13.

14.

15.

16.

17. 18.

19. 20. 21.

22.

23.

24.

25.

on the concentration and metabolism of plasma Iipoproteins. N EnglJ Med 1991;325:1196-204. Sherwin BB, Gelfand MM. A prospective one-year study of estrogen and progestin in postmenopausal women: effects on clinical symptoms and lipoprotein lipids. Obstet Gynecol 1989;73:759-66. Ottosson Uß,Johansson BG, von Schoultz B. Subfractions of high-density lipoprotein cholesterol during estrogen replacement therapy: a comparison between progestogens and natural progesterone. AM J OBSTET GYNECOL1985;151: 746-50. Hirvonen E, Lipasti A, Mälkönen M, et al. Clinical and lipid metabolic effects of unopposed oestrogen and two oestrogen-progestogen regimens in post-menopausal women. Maturitas 1987;9:69-79. Sonnendrecker EWW, Polakow ES, Benade AJS, Simchowitz E. Serum lipoprotein effects of conjugated estrogen and a sequential conjugated estrogen-medrogestone regimen in hysterectomized postmenopausal women. AM J OBSSET GvNE«OI«1989;160:1128-34. Miller VT, Muesing RA, LaRosa JC, Stoy DB, Phillips EA, Stillman RJ. Effects of conjugated equine estrogen with and without three different progestogens on lipoproteins, highdensity lipoprotein subfractions, and apolipoprotein A-I. Obstet Gynecol 1991;77:235-40. The ~¥iting Group for the Postmenopausal Estrogen/Progestin Interventions Trial. Effects of estrogen or estrogen/progestin regimens on heart disease risk factors in postmenopausal women: the Postmenopausal Estrogen/Progestin Interventions (PEPI) Trial. JAMA 1995;273: 199-208. Krauss RM, Lindgren FT, WingerdJ, Bradley DD, Ramcharan S. Effects of estrogens and progestins on high density lipoproteins. Lipids 1979; 14:113-8. Fahraeus L, Wallentin L. High density lipoprotein subfractions during oral and cutaneous administration of 17[3-estradiol to menopausal women. J Clin Endocrinol Metab 1983; 56:797-801. CauteyJA, LePorte RE, Kuller LH, et al. Menopausal esu'ogen use, high density lipoprotein cholesterol subfractions and liver function. Atherosclerosis 1983;49:31-9. Hart DM, Farish E, Fletcher CD, Howie C, Kitchener H. Ten years post-menopausal hormone replacement therapy--effect on lipoproteins. Maturitas 1984;5:271-6. Blumenfeld Z, Aviram M, Brook GJ, BrandesJM. Changes in lipoproteins and subfractions following oophorectomy and oestrogen replacement in peri-menopausal women. Maturitas 1983;5:77-83. Farish E, Fletcher CD, Hart DM, Teo HTHTC, Alazzawi F, Howie C. The effects of conjugated equine oestrogens with and without a cyclical progestogen on lipoproteins, and HDL subfractions in postmenopausal women. Acta Endocrinol (Copenh) 1986;113:123-7. Williams PT, Vranizan KM, Austin MA, Krauss RM. Associations of age, adiposity, alcohol intake, menstrual stares and estrogen therapy with high-density lipoprotein subclasses. Arterioscler Thromb 1993; 13:1654-61. Cheung MC, Brown BG, WolfAC, Albers..[]. Altered particle size distribudon of apolipoprotein A-I-containing lipoproteins in subjects with coronary artery disease. J Lipid Res 1991 ;32:383-94. Barrett-Connor E, Wingard DL, Criqui MH. Postmenopausal estrogen use and heart disease risk factors in the 1980s: Rancho Bernardo, Calif, revisited. JAMA 1989;261: 2095-100.